The dichotomy between the challenging nature of obtaining annotations for activities, and the more straightforward nature of data collection from wearables, has resulted in significant interest in the development of techniques that utilize large quantities of unlabeled data for learning representations. Contrastive Predictive Coding (CPC) is one such method, learning effective representations by leveraging properties of time-series data to setup a contrastive future timestep prediction task. In this work, we propose enhancements to CPC, by systematically investigating the encoder architecture, the aggregator network, and the future timestep prediction, resulting in a fully convolutional architecture, thereby improving parallelizability. Across sensor positions and activities, our method shows substantial improvements on four of six target datasets, demonstrating its ability to empower a wide range of application scenarios. Further, in the presence of very limited labeled data, our technique significantly outperforms both supervised and self-supervised baselines, positively impacting situations where collecting only a few seconds of labeled data may be possible. This is promising, as CPC does not require specialized data transformations or reconstructions for learning effective representations.
Economy is severely dependent on the stock market. An uptrend usually corresponds to prosperity while a downtrend correlates to recession. Predicting the stock market has thus been a centre of research and experiment for a long time. Being able to predict short term movements in the market enables investors to reap greater returns on their investments. Stock prices are extremely volatile and sensitive to financial market. In this paper we use Deep Learning networks to predict stock prices, assimilating financial, business and technology news articles which present information about the market. First, we create a simple Multilayer Perceptron (MLP) network and then expand into more complex Recurrent Neural Network (RNN) like Long Short Term Memory (LSTM), and finally propose FinBERT-LSTM model, which integrates news article sentiments to predict stock price with greater accuracy by analysing short-term market information. We then train the model on NASDAQ-100 index stock data and New York Times news articles to evaluate the performance of MLP, LSTM, FinBERT-LSTM models using mean absolute error (MAE), mean absolute percentage error (MAPE) and accuracy metrics.
Practical operations of coordinated fleets of mobile robots in different environments reveal benefits of maintaining small distances between robots as they move at higher speeds. This is counter-intuitive in that as speed increases, increased distances would give robots a larger time to respond to sudden motion variations in surrounding robots. However, there is a desire to have lower inter-robot distances in examples like autonomous trucks on highways to optimize energy by vehicle drafting or smaller robots in cluttered environments to maintain communication, etc. This work introduces a model based control framework that directly takes non-linear system dynamics into account. Each robot is able to follow closer at high speeds because it makes predictions on the state information from its adjacent robots and biases it's response by anticipating adjacent robots' motion. In contrast to existing controllers, our non-linear model based predictive decentralized controller is able to achieve lower inter-robot distances at higher speeds. We demonstrate the success of our approach through simulated and hardware results on mobile ground robots.
Entity alignment aims to identify equivalent entity pairs between different knowledge graphs (KGs). Recently, the availability of temporal KGs (TKGs) that contain time information created the need for reasoning over time in such TKGs. Existing embedding-based entity alignment approaches disregard time information that commonly exists in many large-scale KGs, leaving much room for improvement. In this paper, we focus on the task of aligning entity pairs between TKGs and propose a novel Time-aware Entity Alignment approach based on Graph Neural Networks (TEA-GNN). We embed entities, relations and timestamps of different KGs into a vector space and use GNNs to learn entity representations. To incorporate both relation and time information into the GNN structure of our model, we use a time-aware attention mechanism which assigns different weights to different nodes with orthogonal transformation matrices computed from embeddings of the relevant relations and timestamps in a neighborhood. Experimental results on multiple real-world TKG datasets show that our method significantly outperforms the state-of-the-art methods due to the inclusion of time information.
Visual Question Answering (VQA) is a multi-discipline research task. To produce the right answer, it requires an understanding of the visual content of images, the natural language questions, as well as commonsense reasoning over the information contained in the image and world knowledge. Recently, large-scale Vision-and-Language Pre-trained Models (VLPMs) have been the mainstream approach to VQA tasks due to their superior performance. The standard practice is to fine-tune large-scale VLPMs pre-trained on huge general-domain datasets using the domain-specific VQA datasets. However, in reality, the application domain can change over time, necessitating VLPMs to continually learn and adapt to new domains without forgetting previously acquired knowledge. Most existing continual learning (CL) research concentrates on unimodal tasks, whereas a more practical application scenario, i.e, CL on cross-domain VQA, has not been studied. Motivated by this, we introduce CL-CrossVQA, a rigorous Continual Learning benchmark for Cross-domain Visual Question Answering, through which we conduct extensive experiments on 4 VLPMs, 4 CL approaches, and 5 VQA datasets from different domains. In addition, by probing the forgetting phenomenon of the intermediate layers, we provide insights into how model architecture affects CL performance, why CL approaches can help mitigate forgetting in VLPMs to some extent, and how to design CL approaches suitable for VLPMs in this challenging continual learning environment. To facilitate future work on CL for cross-domain VQA, we will release our datasets and code.
For node classification, Graph Neural Networks (GNN) assign predefined labels to graph nodes according to node features propagated along the graph structure. Apart from the traditional end-to-end manner inherited from deep learning, many subsequent works input assigned labels into GNNs to improve their classification performance. Such label-inputted GNNs (LGNN) combine the advantages of learnable feature propagation and long-range label propagation, producing state-of-the-art performance on various benchmarks. However, the theoretical foundations of LGNNs are not well-established, and the combination is with seam because the long-range propagation is memory-consuming for optimization. To this end, this work interprets LGNNs with the theory of Implicit GNN (IGNN), which outputs a fixed state point of iterating its network infinite times and optimizes the infinite-range propagation with constant memory consumption. Besides, previous contributions to LGNNs inspire us to overcome the heavy computation in training IGNN by iterating the network only once but starting from historical states, which are randomly masked in forward-pass to implicitly guarantee the existence and uniqueness of the fixed point. Our improvements to IGNNs are network agnostic: for the first time, they are extended with complex networks and applied to large-scale graphs. Experiments on two synthetic and six real-world datasets verify the advantages of our method in terms of long-range dependencies capturing, label transitions modelling, accuracy, scalability, efficiency, and well-posedness.
This paper proposes a Decentralized Stochastic Gradient Descent (DSGD) algorithm to solve distributed machine-learning tasks over wirelessly-connected systems, without the coordination of a base station. It combines local stochastic gradient descent steps with a Non-Coherent Over-The-Air (NCOTA) consensus scheme at the receivers, that enables concurrent transmissions by leveraging the waveform superposition properties of the wireless channels. With NCOTA, local optimization signals are mapped to a mixture of orthogonal preamble sequences and transmitted concurrently over the wireless channel under half-duplex constraints. Consensus is estimated by non-coherently combining the received signals with the preamble sequences and mitigating the impact of noise and fading via a consensus stepsize. NCOTA-DSGD operates without channel state information (typically used in over-the-air computation schemes for channel inversion) and leverages the channel pathloss to mix signals, without explicit knowledge of the mixing weights (typically known in consensus-based optimization). It is shown that, with a suitable tuning of decreasing consensus and learning stepsizes, the error (measured as Euclidean distance) between the local and globally optimum models vanishes with rate $\mathcal O(k^{-1/4})$ after $k$ iterations. NCOTA-DSGD is evaluated numerically by solving an image classification task on the MNIST dataset, cast as a regularized cross-entropy loss minimization. Numerical results depict faster convergence vis-\`a-vis running time than implementations of the classical DSGD algorithm over digital and analog orthogonal channels, when the number of learning devices is large, under stringent delay constraints.
Question answering over temporal knowledge graphs (KGs) efficiently uses facts contained in a temporal KG, which records entity relations and when they occur in time, to answer natural language questions (e.g., "Who was the president of the US before Obama?"). These questions often involve three time-related challenges that previous work fail to adequately address: 1) questions often do not specify exact timestamps of interest (e.g., "Obama" instead of 2000); 2) subtle lexical differences in time relations (e.g., "before" vs "after"); 3) off-the-shelf temporal KG embeddings that previous work builds on ignore the temporal order of timestamps, which is crucial for answering temporal-order related questions. In this paper, we propose a time-sensitive question answering (TSQA) framework to tackle these problems. TSQA features a timestamp estimation module to infer the unwritten timestamp from the question. We also employ a time-sensitive KG encoder to inject ordering information into the temporal KG embeddings that TSQA is based on. With the help of techniques to reduce the search space for potential answers, TSQA significantly outperforms the previous state of the art on a new benchmark for question answering over temporal KGs, especially achieving a 32% (absolute) error reduction on complex questions that require multiple steps of reasoning over facts in the temporal KG.
For Lifelong SLAM, one has to deal with temporary localization failures, e.g., induced by kidnapping. We achieve this by starting a new map and merging it with the previous map as soon as relocalization succeeds. Since relocalization methods are fallible, it can happen that such a merge is invalid, e.g., due to perceptual aliasing. To address this issue, we propose methods to detect and undo invalid merges. These methods compare incoming scans with scans that were previously merged into the current map and consider how well they agree with each other. Evaluation of our methods takes place using a dataset that consists of multiple flat and office environments, as well as the public MIT Stata Center dataset. We show that methods based on a change detection algorithm and on comparison of gridmaps perform well in both environments and can be run in real-time with a reasonable computational cost.
We introduce a Transformer based 6D Object Pose Estimation framework VideoPose, comprising an end-to-end attention based modelling architecture, that attends to previous frames in order to estimate accurate 6D Object Poses in videos. Our approach leverages the temporal information from a video sequence for pose refinement, along with being computationally efficient and robust. Compared to existing methods, our architecture is able to capture and reason from long-range dependencies efficiently, thus iteratively refining over video sequences. Experimental evaluation on the YCB-Video dataset shows that our approach is on par with the state-of-the-art Transformer methods, and performs significantly better relative to CNN based approaches. Further, with a speed of 33 fps, it is also more efficient and therefore applicable to a variety of applications that require real-time object pose estimation. Training code and pretrained models are available at https://github.com/ApoorvaBeedu/VideoPose